Network Working Group                                         D. Crocker
Request for Comments: 577                                       UCLA-NMC
NIC: 19356                                                  October 1973
References: RFC 524, 539, 555

                             Mail Priority

   In RFC 539 (NIC--17644,3d:gy) Postel and I suggested that mail
   senders be allowed to assign a degree of priority to their mail.
   White (RFC 555--17993,6c:gy) objected to defining shades of urgency,
   without having their effects upon the Mail Protocol server also
   defined.

   If priority levels were to be assigned by automata, I would agree
   with Jim.  Unfortunately, the human sender of the mail will usually
   be the one to assign the priority, and humans will not be consistent
   in that assignment.

   Also unfortunately, the concept of urgency is an integral part of
   communication.  If it weren't, we could ignore its inclusion into the
   MP.

   Since distinctions in urgency are useful (necessary?) and since
   humans will be the ones assigning specific degrees of urgency
   (thereby making it impossible for server processes to automatically
   do the "right thing" in response), we suggested only including the
   INFORMATION as part of the protocol.  Let the human and server-
   process receivers decide between themselves how the server-process
   should deal with that information.

   Now that I have argued all that, let me suggest interpretations for
   urgency values.  This is so that programmers can have automata-
   generated mail (e.g., notification of the status of previously sent
   mail) carry reasonable urgency values:

      10  Phone in the middle of the night, if necessary.
       9
       8  Deliver to user's terminal NOW.
       7
       6  Deliver to user's terminal only if user is at "exec"
          level.
       5
       4  Deliver immediately after sign-on or before sign-off.
       3
       2  Deliver into standard mailbox.
       1
       0  Junk Mail

         [ This RFC was put into machine readable form for entry ]
         [ into the online RFC archives by  Martin Lyngvig 7/99 ]